=== phoenix24_ [i=oqlrqqj@ns37986.ovh.net] has joined #upstart === Amaranth [n=travis@ubuntu/member/Amaranth] has joined #upstart === cort [n=sam@62-31-146-25.cable.ubr12.azte.blueyonder.co.uk] has left #upstart ["SEE] === Amaranth_ [n=travis@ubuntu/member/Amaranth] has joined #upstart === phoenix24 [i=powgtz@ns37986.ovh.net] has joined #upstart === Amaranth [n=travis@ubuntu/member/Amaranth] has joined #upstart === phoenix24 [i=glwkifx@ns37986.ovh.net] has joined #upstart === Md [i=md@freenode/staff/md] has joined #upstart === Keybuk [n=scott@quest.netsplit.com] has joined #upstart === mbiebl [i=michael@conference/debconf/x-9f6f1c397de809bb] has joined #upstart [11:41] rah [11:42] Straw Poll: [11:42] a) "start on EVENT and EVENT" / "stop on EVENT or EVENT" [11:42] or [11:42] b) "from EVENT and EVENT" / "until EVENT or EVENT" [12:29] Keybuk: Well, b) more clearly conveys the fact that upstart manages the job and doesn't just start and stop it. === Keybuk hugs nih_tree_*() === mbiebl [i=michael@conference/debconf/x-c38ff726d7fd6bd2] has joined #upstart [01:23] shawarma: from/until fits in better with adding "while" as well [01:23] from starting dbus [01:23] until stopped dbus [01:23] while filesystem-writable and network-available [01:24] (other suggestions from the floor welcome ) === sadleder [n=sadleder@p5081474A.dip0.t-ipconnect.de] has joined #upstart === sadleder [n=sadleder@p5081474A.dip0.t-ipconnect.de] has left #upstart [] === Starting logfile irclogs/upstart.log === ubuntulog [i=ubuntulo@ubuntu/bot/ubuntulog] has joined #upstart === Topic for #upstart: Upstart 0.3.8 | http://upstart.ubuntu.com/ | http://upstart.ubuntu.com/wiki/UpstartOnGentoo === Topic (#upstart): set by Md at Sun May 6 19:59:46 2007 === Trevelyan` [n=Jim@unaffiliated/trevelyan] has joined #upstart [03:55] hi. looking at launchpad it seems that init script replacement jobs writting seems to have stalled. is this true or has the work moved elsewhere? [03:56] Trevelyan`: we're working on changes to Upstart itself to make it possible [03:58] still not happy with how jobs (defs) work? [03:59] largely, yes [03:59] the issue has been the implementation of more complex event expressions than just an "OR" list [04:00] Upstart 0.2 and 0.3 just implement "start on"/"stop on" as a list of events, and any can match [04:00] that's obviously not flexible enough (and we knew it wouldn't be) [04:00] the delay has been coming up with a replacement expression language that is as flexible as we want, but with very well defined behaviour [04:00] I think we're on a very good track for it now, with some really nice features [04:04] cool. I tried initng, but upstart i think has more promise. however until it has jobs that will do parallel job/process starts i wont get much use out of it, ifykwim. [04:04] thank you for explaining that. [04:05] these things take some amount of time to get right [04:05] no doubt [04:05] and I'm being deliberately cautious, rolling out changes incrementally and testing them, rather than trying to do it all at once [04:06] sensible, but having said that it surprises me more that upstart went into ubuntu so quick. [04:06] how so? [04:07] in order to test, you need a group of users to test it on [04:07] not that it wasn't ready, but how young it was [04:07] that is true === mbiebl [i=michael@conference/debconf/x-a07c00e6b6d16ce5] has joined #upstart [04:09] the set of features of 0.2.7 was a superset of those of sysvinit [04:09] so it was an ideal way to test it, since it could behave like sysvinit while stress-testing the new code in the field [04:11] i gather it went well? since i've not really heard anything that negative about the upstart in ubuntu. [04:13] anyway i look forward to upstart reaching its full potential, and my laptop booting in a couple of seconds =) [04:13] it's gone well, yes [04:14] it's taking a little longer to develop than anticipated, but there's no major panics yet :) [04:15] trevelyan: While youre at it, replace the BIOS with something better and faster. :-) [04:16] http://linuxbios.org/index.php/Main_Page [04:16] Yes. :-) [04:18] meh, BIOS isn't that slow [04:18] although i think that might be a step too far for Dell's tech support to swallow =) [04:18] they only just about accept the fact we run debian [04:20] although since they now ship ubuntu i guess they can't blink at our ubuntu desktops. [04:24] i don't know what dell uses but for their desktops/laptops it fairly fast on what i like to think as cached boots (ie nothings hw or bios wise has changed since last boot) [04:24] on their servers it can take as long as 5 mins before you reach grub =/ === juergbi_ [n=juerg@80-219-19-116.dclient.hispeed.ch] has joined #upstart === mbiebl [i=michael@conference/debconf/x-5a099aaeee12455a] has joined #upstart === Amaranth [n=travis@ubuntu/member/Amaranth] has joined #upstart === mbiebl_ [i=michael@conference/debconf/x-781a73d31aa734fd] has joined #upstart [06:07] Keybuk: what would i need to do if i wanted to add a new command to initctl to change profile? [06:09] AlexExtreme: it's a little complicated ;) [06:09] heh [06:09] first you need to modify libupstart [06:09] add appropriate messages and replies to upstart/message.h, and then modify upstart/message.c so that the arguments in both directions are correctly marshalled [06:10] the usual format here is a "change profile" message from initctl, along with a generic "current profile" message that can be used as both a reply, and as a subscription event [06:10] cf. the emit/event series [06:10] second you need to modify upstart itself [06:11] so that it receives the message and does the appropriate thing with it? [06:11] this is quite easy, just add a function to init/control.c and hook it into the message table, that handles your marshalled message - it looks like an ordinary function that has some useful effect [06:11] and lastly you need to modify initctl [06:11] add an action function that makes the appropriate initctl_send call [06:12] and add handle functions to deal with the replies [06:12] ok [06:15] http://upstart.ubuntu.com/wiki/JobEventExpressions [06:15] ^ for proof-reading and comment [06:15] *clicks* === mbiebl_ [i=michael@conference/debconf/x-ac2c41e20755d2c7] has joined #upstart [06:17] there's a small typo in the design section [06:17] "The expression, or sub-expressions within parentheses) must" [06:17] where's the opening bracket for that closing bracket? ;) [06:18] heh [06:18] fix it for me :p [06:18] k :P [06:19] done [06:19] any comments on the design, rather than my typing? :p [06:19] i haven't finished reading yet ;) [06:19] bah, what is it with the weather in this country? [06:19] this is supposed to be summer! [06:20] gah, thunderstorm overhead, i'll bbiab [06:21] it's doing similar here [06:40] Keybuk, I don't see our "from" idea there. Has that cahnged? [06:40] oh wait there it is. [06:40] unresolved issues heh [06:41] :) [06:54] back === AlexExtreme continues reading the spec [07:02] Keybuk: IMO the from/until is better because it makes it reading a job sound more like the time period in which the job is run (i hope that makes sense :)) === mbiebl [i=michael@conference/debconf/x-8ea951b7c2f4f08a] has joined #upstart [07:08] this seems like a "simpler" verson of c-e-c === mbiebl [i=michael@conference/debconf/x-40b85494925c0240] has joined #upstart === mbiebl [i=michael@conference/debconf/x-db21024cb578a97b] has joined #upstart [07:36] AlexExtreme: it's intended to be step #1 in replacing c-e-c [07:36] step #2 is to introduce "States" which are jobs without the jobbish bits. ie just from/until [07:37] step #3 is to allow expansion of from events in the until expression [07:37] step # [07:37] states introduce a "while" stanza [07:37] and I do think that from/until/while go together nicely [07:40] cool, sounds good === Md [i=md@freenode/staff/md] has joined #upstart === juergbi [n=juerg@80-219-17-66.dclient.hispeed.ch] has joined #upstart === juergbi [n=juerg@80-219-17-66.dclient.hispeed.ch] has joined #upstart === Md [i=md@freenode/staff/md] has joined #upstart === Trevelyan` [n=Jim@unaffiliated/trevelyan] has joined #upstart === tale [n=tale@207.235.54.1] has joined #upstart === Keybuk [n=scott@quest.netsplit.com] has joined #upstart === ion_ [i=ion@heh.fi] has joined #upstart === kylem [n=kyle@amnesiac.heapspace.net] has joined #upstart === thom [n=thom@amnesiac.heapspace.net] has joined #upstart === dg [i=dgl@otherwize.co.uk] has joined #upstart === shawarma [n=sh@atlas.linux2go.dk] has joined #upstart === jonib1 [n=jonas@213.212.2.165] has joined #upstart === AlexExtreme [n=AlexExtr@frugalware/developer/AlexExtreme] has joined #upstart === crazy [n=crazy@frugalware/developer/crazy] has joined #upstart === benmur [n=benmur@friends.sukria.net] has joined #upstart === dAndy [i=dandy@serenity.cat.pdx.edu] has joined #upstart === wasabi [n=wasabi@207.55.182.70] has joined #upstart === Amaranth [n=travis@ubuntu/member/Amaranth] has joined #upstart === phoenix24 [i=ahqub@ns37986.ovh.net] has joined #upstart [09:48] Wow. By using the GUI edit mode, i managed to do a huge diff, while adding just a single character. http://upstart.ubuntu.com/wiki/JobEventExpressions?action=diff&rev2=3&rev1=2 [09:48] It changed each `...` to {{{...}}} [09:54] Is it possible to subscribe to each current and future spec at the Upstart wiki? === phoenix24_ [i=rqf@ns37986.ovh.net] has joined #upstart